Competition activity information display method and device, equipment, medium and product

Abstract

The invention discloses an event information display method, device and equipment, a medium and a product, and the method comprises the steps: predicting obtained historical sample data according to a student model in a knowledge distillation model obtained through pre-training, and obtaining key factors of an event; predicting the historical sample data according to a teacher model in the knowledge distillation model to obtain a prediction result of the match; and displaying and outputting the key factor and the prediction result as prediction. According to the scheme, the key factor and the prediction result are used as prediction display output, the explanation copywriting is automatically generated and displayed to the user in combination with personalized features of the user, a more comprehensive and specific prediction analysis result is provided for the user, the accuracy of content recommendation is improved, and the richness of prediction information is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, equipment, medium and product for displaying event information. Background Art

[0002] Sports fans and data analysts have become increasingly entrenched in sports analysis, especially in predicting the results of games before and during the games. Users use game prediction solutions to understand the results of games they are interested in, as a basis for their choice of games to watch.

[0003] However, when predicting existing events, for each game, it is necessary to invite experts in the football field to interpret the experience and output the prediction results of the game. This solution is costly, the prediction results are more subjective, the accuracy is low, and it is not possible to output the user's interests in a targeted manner according to the user's personalized needs. Another solution is to use artificial intelligence to determine the prediction results. This solution is more efficient and convenient, but it can only achieve result prediction, and cannot make a reasonable explanation for the result prediction, and the prediction information provided to users is single. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a method, device, equipment, medium and product for displaying event information to improve the richness of prediction information.

[0005] An embodiment of the present invention provides a method for displaying event information, the method comprising:

[0006] The student model in the pre-trained knowledge distillation model is used to predict the acquired historical sample data to obtain the key factors of the event;

[0007] Predicting the historical sample data according to the teacher model in the knowledge distillation model to obtain a prediction result of the event;

[0008] The key factors and the prediction results are output as a prediction display.

[0009] Preferably, the key factors and the prediction results are output as prediction display, including:

[0010] When the user enters the prediction list page, a prediction list is generated and output based on the key factors of each event;

[0011] When the user enters the detail list page of the event to be displayed, a detail page is generated and output according to the prediction results of the event to be displayed.

[0012] As a preferred embodiment, the method further comprises:

[0013] Calculate the influence weights of different feature values ​​in the historical sample data according to the weight parameters of the student model obtained by joint training of the knowledge distillation model;

[0014] The characteristic value serving as the key factor is determined according to the influence weight.

[0015] Preferably, a prediction list is generated and output based on the key factors of each event, including:

[0016] Get the historical data of the user watching the event;

[0017] Count the number of times users have viewed each team, and count the number of times users have viewed each contestant based on the relationship between the team and the contestants;

[0018] Generate the user's preference for each team and each contestant based on the number of times the user has viewed each team and the number of times the user has viewed each contestant;

[0019] The prediction list is generated and outputted according to the key factors of each event, the user's preference for each team, and the user's preference for each participant.

[0020] Preferably, the prediction list is generated and outputted according to the key factors of each event, the user's preference for each team, and the user's preference for each participant, including:

[0021] Determine a list of recommended events for the user based on the user's preference for each team and each participant;

[0022] According to the maximum weight factor among the key factors of each recommended event, characteristic data of both participants in each recommended event are obtained;

[0023] The advantages and disadvantages of the characteristic data of the two contestants of each recommended event are compared, and the explanation copy of each recommended event is output to obtain the prediction list.

[0024] As a preferred solution, the knowledge distillation model training process specifically includes:

[0025] Define teacher model and student model;

[0026] Define the loss function for model training;

[0027] The loss function is minimized using the gradient descent method, and the model parameters of the teacher model and the student model are updated to obtain the knowledge distillation model.

[0028] The embodiment of the present invention further provides a device for displaying event information, the device comprising:

[0029] The explanation module is used to predict the acquired historical sample data based on the student model in the pre-trained knowledge distillation model to obtain the key factors of the event;

[0030] A result module is used to predict the historical sample data according to the teacher model in the knowledge distillation model to obtain the prediction result of the event;

[0031] A display module is used to output the key factors and the prediction results as prediction displays.

[0032] An embodiment of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, a method for displaying event information as described in any of the above embodiments is implemented.

[0033] An embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a method for displaying event information as described in any of the above embodiments.

[0034] An embodiment of the present invention further provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any of the above embodiments when executed by a processor.

[0035] Compared with the prior art, the present invention provides a method, device, equipment, medium and product for displaying event information. The student model in the pre-trained knowledge distillation model is used to predict the acquired historical sample data to obtain the key factors of the event; the teacher model in the knowledge distillation model is used to predict the historical sample data to obtain the prediction results of the event; and the key factors and the prediction results are output as prediction display. The present application scheme outputs the key factors and the prediction results as prediction display, automatically generates explanatory text, and displays it to the user in combination with the user's personalized characteristics, providing the user with more comprehensive and specific prediction analysis results, improving the accuracy of content recommendations, and increasing the richness of prediction information. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flowchart of a method for displaying event information provided by an embodiment of the present invention;

[0037] Figure 2 is another flow chart of the method for displaying event information provided by an embodiment of the present invention;

[0038] Figure 3It is a structural schematic diagram of a device for displaying event information provided by an embodiment of the present invention;

[0039] Figure 4 It is a structural diagram of a terminal device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] It can be understood that the various embodiments described below can be combined or combined in a logical manner, and the embodiments of the present invention will not describe various combinations one by one.

[0042] This application proposes a method for displaying event information. Figure 1 , is a flow chart of a method for displaying event information provided by an embodiment of the present invention, the method comprising the following steps:

[0043] Step S1, predicting the acquired historical sample data according to the student model in the pre-trained knowledge distillation model to obtain the key factors of the event;

[0044] Step S2, predicting the historical sample data according to the teacher model in the knowledge distillation model to obtain a prediction result of the event;

[0045] Step S3: output the key factors and the prediction results as prediction display.

[0046] In the specific implementation of this embodiment, the prediction of events is basically based on the manual experience of experts in the field, which is costly due to the reliance on expert experience. Especially when the number of events is large and the types are various, the cost of relying on expert experience to make judgments will increase exponentially.

[0047] Among the existing solutions, there is a solution that uses a deep neural network model to predict the results of an event. The deep neural network model performs well in terms of model prediction accuracy due to its rich data expression capabilities. However, due to the complex structure of the model, the prediction results are poorly interpretable, that is, it is impossible to provide a basis for specific prediction analysis and prediction influencing factors. When users actually refer to the prediction results, they are not only concerned with the predicted event results. These explanatory factors are actually the information that users should pay attention to, which affects users' predictions for the event and their choices of watching the event. Since the prediction models of the prior art lack the key factors for prediction explanation, in response to the above technical problems, the method for displaying event information improved in this case includes the following steps when it is specifically implemented:

[0048] Obtain historical sample data for users to conduct predictive analysis. The historical sample data specifically includes historical match data of both teams in the match and feature data that affects the probability of a team winning.

[0049] The student model in the pre-trained knowledge distillation model is used to predict the historical sample data to obtain the key factors of the event; the key factors of the event are the key factors that affect the game. The key factors are used to determine the event explanation copy and provide more display information to users for reference.

[0050] The historical sample data is predicted according to the teacher model in the knowledge distillation model to obtain the prediction result of the event, which is the winning probability of the two participating teams.

[0051] The knowledge distillation model is a machine learning technique that trains a small student model to mimic the behavior of a large teacher model. During model training, the teacher model and the student model are trained jointly, and the student model can be used alone for prediction and reasoning. This approach helps reduce the size of the model, speeds up reasoning, and is used to improve the efficiency and scalability of the model while maintaining prediction accuracy.

[0052] The event prediction model first needs to be accurate enough, which is the core competitiveness of the entire solution. Therefore, when predicting the results in this case, the teacher model needs to be used to predict the results to ensure the prediction accuracy. At the same time, in order to show users the important factors that affect the game, the model needs to be sufficiently interpretable, so the student model of the knowledge distillation model is used to complete this task.

[0053] The deep neural network model has excellent performance in model prediction accuracy due to its rich data expression ability, but the model structure is complex and the interpretability of the prediction results is poor; the logistic regression model based on machine learning has poor data expression ability due to its simple structure and is not as accurate as the deep neural network, but the interpretability is good enough to facilitate the extraction of key factors affecting the prediction results. In order to take both into account, this proposal proposes a method of using knowledge distillation combined with model training to predict the results of football games.

[0054] Knowledge distillation joint model training can ensure the consistency of deep neural network and logistic regression model training objectives; the deep neural network model as the teacher model is responsible for predicting the results of the game and improving the accuracy of the prediction results; the logistic regression model as the student model is easy to interpret and can be used to extract the maximum weight factor that affects the prediction results and construct personalized copywriting for downstream tasks.

[0055] The key factors and the prediction results are output as prediction displays, and explanation text is automatically generated. These are displayed to users in combination with their personalized characteristics, providing users with more comprehensive and specific prediction analysis results for their reference, thereby improving the richness of prediction information and facilitating users to refer to and decide on their preferences for watching events.

[0056] In another embodiment provided by the present invention, see Figure 2 , is another flow chart of the method for displaying event information provided by an embodiment of the present invention.

[0057] In the specific implementation, the knowledge distillation model is constructed in advance using joint distillation learning. The logistic regression model used as the student model in the knowledge distillation model is used to analyze the historical competition data samples of the competition teams to determine the student result prediction, which is the key factor of the competition.

[0058] When the first signal triggered by the user is received, the prediction list page is entered, and a prediction list is generated and output according to the key factors of each event as the prediction list page data of the game, showing the key factors of each event to the user for reference.

[0059] The deep neural network used as the teacher model in the knowledge distillation model is used to analyze the historical game data samples of the competition teams to determine the teaser result prediction, that is, the game prediction result;

[0060] When the second signal triggered by the user is received, the prediction details page of the event to be displayed is entered, and a prediction list is generated and output according to the game prediction results of the event to be displayed, as the prediction details page of the game, to display the event prediction results of the event to be displayed to the user.

[0061] It should be noted that the prediction display data scheme provided in this case is only a preferred implementation method. In other embodiments, other schemes can be used to display the key factors and prediction results of the prediction. For example, only the key factors with the highest priority are displayed on the list page, while the event prediction results and more comprehensive event key factors are displayed on the prediction details page, so as to provide users with more comprehensive reference information.

[0062] Based on the logistic regression model and deep neural network model in joint distillation learning, different display pages are designed to display different information to users. The list page provides users with the analysis results of key factors to facilitate users to choose specific events to pay attention to. The details page provides specific prediction results to attract users to click and interact, improve the interactivity of event information display, and enhance user experience.

[0063] In another embodiment provided by the present invention, the logistic regression model analyzes the historical match data samples of the event teams to determine the key factors of the event. When a football match prediction scheme needs to be generated, while using the deep neural network to predict the match results, the jointly trained logistic regression model is used to find the maximum weight factor. The process is as follows:

[0064] Using the trained parameters of the logistic regression model, we can extract the maximum weight factor that affects the prediction results. The following formula is the formula of the logistic regression model:

[0065] The weight parameters of the student model calculate different characteristic values ​​in the historical sample data to extract the maximum weight factor affecting the prediction result;

[0066] Among them, p is the predicted probability. If it is greater than 0.5, it means that the home team will win, and if it is less than 0.5, it means that the home team will lose. The factors affecting p are composed of a series of w and x. Among them, w is the trained parameter, and x is the input feature value. In the column formula, the largest w*x is the maximum weight factor, and w*x can be interpreted as:

[0067] "The difference in winning rate between the two teams in the last 10 games is x = 0.2, and the model parameter w = 0.1."

[0068] "The points difference between the two teams in the last 10 games is x = 0.8 (normalized), and the model parameter w = 0.23."

[0069] "The goal difference between the two teams in the last 10 games is x=0.3 (normalized), and the model parameter w=0.2".

[0070] Then we can determine that the biggest factor affecting this game is "the points difference between the two teams in the last 10 games" (because 0.8*0.23>0.3*0.2>0.2*0.1).

[0071] Therefore, the product of different eigenvalues ​​of the model parameters and the model parameters is used as the influence weight, and the eigenvalue seat with the largest product is taken as the maximum weight factor affecting the prediction result, which is taken as the key factor.

[0072] It should be noted that the number of key factors can be determined according to the number of key factors that need to be selected.

[0073] By using the logistic regression model as a student model to analyze the features, we can determine the degree of influence of different feature values, and then determine the influencing factors with the greatest influence. Through the influencing factor analysis, we can provide feedback on the influencing factors of the event results, generate corresponding explanatory texts, and display them to users, thereby increasing the richness of the event information display and attracting users.

[0074] In another embodiment provided by the present invention, when generating a prediction list, it is also necessary to comprehensively consider the user's preferences and habits to generate a personalized prediction list.

[0075] Get the historical data of the user watching the event;

[0076] According to the user's history of watching games, count the number of times the user watched team n, t n At the same time, according to the affiliation relationship between the team and the contestants, the number of times users watch contestant i, pi, is counted.

[0077] According to the number of times the user has viewed each team and the number of times the user has viewed each contestant, the user's preference for each team and the user's preference for each contestant are generated.

[0078] Generate the user's preference for two types of features, team and contestant, based on the preference level, i.e., the viewing preference ranking or preference weight of the team obtained based on the number of views of the team and the total number of views of the user;

[0079] The preference degree of the contestants is the viewing preference ranking or preference weight of the contestants obtained based on the viewing times of the contestants and the overall viewing times of the users;

[0080] According to the user's preference for different teams and participants, a preset number of events with the highest preference among all events are determined as list events, and the number of list events is the preset number of list displays.

[0081] The analysis results of key factors of different events are displayed in a list to obtain a prediction list of different events.

[0082] By collecting statistics on the viewing data of different users, personalized prediction lists are output for different users, corresponding prediction analysis information is displayed, and personalized event information display is provided.

[0083] In another embodiment provided by the present invention, when outputting the prediction list, the list of recommended events for the user is determined based on the user's preference for each team and each contestant.

[0084] Specifically, according to the preference levels of the teams and participants, the comprehensive preferences of the two participating teams in the event are calculated according to a preset weighted ratio, that is, the preferences of each team and each participant are calculated to determine the preferences of the event teams, and the sum of the preference levels of the two teams is taken as the comprehensive preference of the event.

[0085] After determining the comprehensive preference of the events, the recommended events with the best preference are determined in the order of the comprehensive preference.

[0086] After determining the recommended events, you also need to determine the corresponding list copy for different events. According to the maximum weight factor among the key factors of each recommended event, obtain the characteristic data of the two participants in each recommended event, that is, according to the maximum weight factor output by the student model, obtain the characteristic data of both parties in the event.

[0087] The feature data is matched to determine the pros and cons of the feature data, and the explanation text of each recommended match is output to obtain the prediction list.

[0088] For example, if the maximum weight factor is generated by the historical battle characteristics; the corresponding generation includes: Copy 1: "You often see that XXX team is worse than its opponents in terms of goals / win rate / points recently. You can unlock the plan to view the details." Copy 2: "Your favorite XXX team is not as good as its opponents in terms of goals / win rate / points recently. Click to unlock to help you analyze the gap."

[0089] If the maximum weight factor is generated by player characteristics, the corresponding generated copy is Copy 1: "Your favorite player XXX has performed well recently. Click to unlock and learn more details."

[0090] If the largest weight factor is the environmental feature - weather, get the current weather forecast. If the weather is consistent, display it first and show the weather icon. "Rainy days will have a significant impact on the outcome of the game."

[0091] If the largest weight factor is the environmental feature - weather, the user's geographic LBS information is obtained. If it matches, the name of the stadium is displayed. This is mainly given priority to users who are watching the game on site.

[0092] It should be noted that the specific problem generation solution is specifically combined with the determined feature data comparison and is generated using the conventional existing technology Chinese copy generation model.

[0093] This case obtains a list of match prediction information that can be displayed, and each match is accompanied by the biggest factor that may affect the outcome of the match.

[0094] Match user characteristics with the factors of each game one by one and generate different copywriting. The copywriting list involved in this proposal is as follows:

[0095] Table 1 Comparison table of personalized copywriting.

[0096]

[0097] From the comparison of the above tables, it can be seen that based on the user's preference statistics and combined with the actual data of the event teams, it is possible to output the copy content that the user is "more interested in" as a predicted list of the event, attracting users to click on specific details and interact, thereby improving the efficiency of displaying event information.

[0098] In another embodiment provided by the present invention, the knowledge distillation model is specifically trained, including the following steps:

[0099] The teacher model T(x) is defined based on a deep neural network, and the student model S(x) is defined based on a logistic regression model.

[0100] Define the loss function for model training.

[0101] The loss function is Loss = H(y, S(x)) + μ*||logits(T(x))-logits(S(x))||.

[0102] Among them, y is the training target, that is, the game is a "win, lose, draw" result, μ is the "temperature control parameter", which indicates how much the influence of the teacher model is adjusted to the student model. Logits(f(x)) represents the value before the normalized output of the last layer of the model network.

[0103] Use the gradient descent method to minimize the Loss function and obtain the teacher model T(x) and student model S(x) after parameter update.

[0104] Through joint distillation learning, the student model is trained to imitate the teacher model, ensuring the efficiency of the student model while improving the accuracy and ensuring the accuracy of the key factors of the output.

[0105] In another embodiment provided by the present invention, the key factors include at least one of participant characteristics, team characteristics, stadium characteristics and weather characteristics.

[0106] The factors that affect the actual game are as follows: player characteristics, including player name and player position; team characteristics, the historical characteristics of the two teams, including the difference in the number of goals, winning rate and points in the last 10 games; stadium characteristics, past experience shows that the atmosphere of the stadium has a greater impact on the outcome of the game; weather characteristics, the good or bad weather has a great impact on the outcome of the game.

[0107] By comprehensively analyzing the characteristic factors that affect the event, the source of the recommendation reasons is determined, ensuring the accuracy of the explanation copy of the recommendation reasons.

[0108] See also Figure 3 , is a schematic diagram of the structure of a device for displaying event information provided by an embodiment of the present invention, the device comprising:

[0109] The explanation module is used to predict the acquired historical sample data based on the student model in the pre-trained knowledge distillation model to obtain the key factors of the event;

[0110] A result module is used to predict the historical sample data according to the teacher model in the knowledge distillation model to obtain the prediction result of the event;

[0111] A display module is used to output the key factors and the prediction results as prediction displays.

[0112] The event information display device provided in this embodiment can execute all steps and functions of the event information display method provided in any of the above embodiments, and the specific functions of the device will not be described in detail here.

[0113] See also Figure 4 , is a schematic diagram of the structure of a terminal device provided in an embodiment of the present invention. The terminal device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a program for displaying event information. When the processor executes the computer program, the steps in each of the above-mentioned embodiments of the method for displaying event information are implemented, such as Figure 1 Alternatively, the processor implements the functions of each module in the above-mentioned device embodiments when executing the computer program.

[0114] Exemplarily, the computer program may be divided into one or more modules, and the one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments that can perform specific functions, and the instruction segments are used to describe the execution process of the computer program in the event information display device. For example, the computer program may be divided into several modules, and the specific functions of each module have been described in detail in a method for displaying event information provided in any of the above embodiments, and the specific functions of the device will not be repeated here.

[0115] The event information display device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The event information display device may include, but is not limited to, a processor and a memory. A person skilled in the art may understand that the schematic diagram is only an example of an event information display device and does not constitute a limitation on an event information display device. It may include more or fewer components than shown in the diagram, or a combination of certain components, or different components. For example, the event information display device may also include input and output devices, network access devices, buses, etc.

[0116] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the event information display device, and uses various interfaces and lines to connect the various parts of the event information display device.

[0117] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the event information display device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0118] Wherein, if the module integrated in the event information display device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0119] An embodiment of the present invention further provides a computer program product, including a computer program / instruction, which implements the steps of any method described in the above embodiments when executed by a processor.

[0120] It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for displaying event information, characterized in that: The method comprises: The student model in the pre-trained knowledge distillation model is used to predict the acquired historical sample data to obtain the key factors of the event; Predicting the historical sample data according to the teacher model in the knowledge distillation model to obtain a prediction result of the event; The key factors and the prediction results are output as a prediction display.

2. The method for displaying event information according to claim 1, characterized in that: Outputting the key factors and the prediction results as prediction display includes: When the user enters the prediction list page, a prediction list is generated and output based on the key factors of each event; When the user enters the detail list page of the event to be displayed, a detail page is generated and output according to the prediction results of the event to be displayed.

3. The method for displaying event information according to claim 1, characterized in that: The method further comprises: Calculate the influence weights of different feature values ​​in the historical sample data according to the weight parameters of the student model obtained by joint training of the knowledge distillation model; The characteristic value serving as the key factor is determined according to the influence weight.

4. The method for displaying event information according to claim 1, characterized in that: Generate and output a prediction list based on key factors for each event, including: Get the historical data of the user watching the event; Count the number of times users have viewed each team, and count the number of times users have viewed each contestant based on the relationship between the team and the contestants; Generate the user's preference for each team and each contestant based on the number of times the user has viewed each team and the number of times the user has viewed each contestant; The prediction list is generated and outputted according to the key factors of each event, the user's preference for each team, and the user's preference for each participant.

5. The method for displaying event information according to claim 4, characterized in that: The prediction list is generated and outputted according to the key factors of each event, the user's preference for each team, and the user's preference for each participant, including: Determine a list of recommended events for the user based on the user's preference for each team and each participant; According to the maximum weight factor among the key factors of each recommended event, characteristic data of both participants in each recommended event are obtained; The advantages and disadvantages of the characteristic data of the two contestants of each recommended event are compared, and the explanation copy of each recommended event is output to obtain the prediction list.

6. The method for displaying event information according to claim 1, characterized in that: The knowledge distillation model training process specifically includes: Define teacher model and student model; Define the loss function for model training; The loss function is minimized using the gradient descent method, and the model parameters of the teacher model and the student model are updated to obtain the knowledge distillation model.

7. A device for displaying event information, characterized in that: The device comprises: The explanation module is used to predict the acquired historical sample data based on the student model in the pre-trained knowledge distillation model to obtain the key factors of the event; A result module is used to predict the historical sample data according to the teacher model in the knowledge distillation model to obtain the prediction result of the event; A display module is used to output the key factors and the prediction results as prediction displays.

8. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for displaying event information as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for displaying event information according to any one of claims 1 to 6.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of any method described in claims 1 to 6 are implemented.

Images